Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~340~12
METHODS FOR IMPROVED TARGETING OF ANTIBODY,
ANTIBODY FRAGMENTS, HORMONES AND OTHER TARGETING AGENTS,
AND CONJUGATES THEREOF
The present invention generally relates to methods for enhancing targeting of
antibodies, antibody fragments, peptide hormones and steroid hormones, and conjugates
thereof. More specifically, methods are disclosed employing blocking antibodies,fragments, hormones and other targeting agents, and conjugates thereof to reduce cross-
10 relative and nonspecific binding of specific antibodies, hormones and other targetingagents to non-target cells.
Antibodies are proteins that have a binding site that is specific for a particular
determin~nt, e.g. antigen or epitope, and other portions that bind to normal tissues in a
15 nonspecific fashion. There are several irnmunological concepts, all related to antibody
binding, that require definition.
T~rget-specific binlling: Binding ofthe antibody, whole or fragment, hormones, other
targeting agents or conjugate thereof, through the antibody's binding site, to the epitope
20 recognized by said antibody on cells expressing said epitope's or hormone's receptor,
where said cells are the desired target of the antibody, whole or fragment, or hormone,
other targeting agent, or conjugates thereof.
An example of target-specific binding is binding of the antibody, whole or fragment,
25 or conjugate thereof, to tumor cells where the antibody in question can also bind
specifically to normal cells. The component of binding to the tumor cells is target-
specific. Another example is binding of bombesin, or gastrin-releasing peptide, to small
cell lung carcinoma.
Cross-reactive bin~ling: Binding of the antibody, whole or fragment, or hormone,other targeting agent or conjugate thereof, through the antibody's binding site, to the
epitope recognized by said antibody on cells ~ essillg said epitope or hormone receptor,
where said cells are not the desired target of the antibody, whole or fragment, or hormone
or conjugate thereof.
X
1~40~t2
-2 -
An example of cross-specific binding is binding of the antibody, whole or fragment,
or conjugate thereof, to normal lung by antibody binding site to the same or structurally
homologous epitope as is resent on a tumor cell. The component of binding to the normal
lung cells by the antigen-binding site of the antibody is cross-reactive. Another example
5 is the binding of bombesin, or gastrin-releasing peptide, to normal cells in the stomach
or pancreas.
Non.~ecific bin(ling: Binding of an antibody, whole or fragment, or hormone or
conjugate thereof, through some mechanism other than the antigen-recognition binding
10 site of the antibody or hormone, to cells other than the target cells.
An example of nonspecific binding is the uptake of antibody into the liver and spleen
due to binding of the antibody by its Fc receptors onto cells in those organs.
A second example would be binding of mannose present in the ricin of antibody-ricin
conjugates to mannose receptors on liver cells.
Specific ~ntihody: Antibody that binds to epitope on desired target cells through its
antigen-recognition sites. Specific antibodies may also bind to epitope or structural
20 homolog present on non-target cells.
Trrelev~nt ~ntibody: Antibody that does not bind to target cells by means of itsantigen-recognition sites, but may bind to non-target and target cells through non-specific
mech~ni~m~, e.g. Fc portion of antibody binding to Fc receptors on cells in
25 reticuloendothelial system (RES).
Rlocking ;~ntibody: Antibody that inhibits the nonspecific binding of
pharmaceutically active specific antibody. Blocking antibodies may include irrelevant
antibody or pharmaceutically inactive specific antibody or fragments or combinations
30 thereof. The latter may also be called "cold-specific" antibody.
,,,
,., 1~4o3l2
-3 -
Ph~rm~ellt~ y active ~ntihody: Antibody that is diagnostically or therapeutically
effective.
The use of antibodies as carriers for radionuclides and cytotoxins has been a goal of
cancer diagnosis and treatment since Pressman et al. showed that l3'I-labelled rabbit anti-
rat kidney antibodies localized in the kidney after intravenous injection (Pressmen, D.,
Keighly, G (1948) J. Tmmllnol 59:141-46). Just a few years later, Vial and Callahan
reported a dramatic, complete response in a patient with widely metastatic malignant
melanoma treated with '3'-I-antibodies raised against his own tumor (Vial, A.B. and
C~ h~n, W. (1956), Univ. Mich. Med. Rllll. 20: 284-86). In the 1960s Bale and co-
workers demonstrated that labelled antibodies to fibrin, which is often deposited in
rapidly growing tumors, could localize in rat, dog and human tumors (65% of 141
patients) (Bale, W.F., et al. (1960) C~ncer Rex. 20: 1501-1504 (1960); McCaidle, R.J.,
Harper, P.V., Spar, I.L., Bale, W.F., Manack, D., et al. (1969) Cancer 1~: 731-59; Bale,
W.F., Centrenia, M.A., Goody, E.D. (1980) C~ncer Res ~L12: 3965-2972). Several years
later, Chao et al. Demonstrated selective uptake of antibody fragments in tumors (Chao,
H.F., Peiper, S.C., Philpott, G.W., et al (1974) Res. Comm ;n Chem P~th & Ph~rm. ~:
749-61).
Monoclonal antibodies (Kohler, G. And Milstein, C. (1975) Nature ~: 495-97)
offer advantages over the polyclonals used in these studies because of their improved
specificity, purity and consistency among lots. These factors, plus their wide availability,
have led to improved clinical applications of antibodies and their conjugates.
Even with monoclonals, however, most antibodies to human tumors have some
normal tissue cross-reactivity. Compared to tumors, these cross-reactive sites may
equally or pl~elelllially bind injected antibody or conjugate and thus absorb a substantial
portion of the ~mini~tered dose, especially if these sites are concentrated in well-
perfused organs. If the antibody is conjugated to a toxic agent, there may be toxicity to
the normal tissues that could be dose-limiting. Therefore, reducing normal tissue binding
of the antibody or conjugates without adversely affecting their tumor localization would
be advantageous.
X
~ 1340312
Also needed in the art are methods to improve the targeting of immunoconjugates to
tumor cells. Most immunoconjugates are produced by chemically linking an antibody to
another agent. Another possibility is creating a fusion protein. The antibody itself, the
process of linking, or the conjugated agent itself may cause decreased localization to the
5 tumor due to nonspecific or cross-reactive binding.
Also needed in the art is a method to improve delivery of cytotoxins or biologicresponse modifiers (BRMs) to tumors using antibodies as carriers, while minimizing
toxicity.
Also needed in the art are similar methods to improve delivery of cytotoxins or
biologic response modifiers (BRMs) to tumors using hormones or other targeting agents
as carriers, while minimi~ing toxicity.
Also needed in the art is a method to decrease formation of antiglobulins to injected
antibody or immunoconjugates as there may be inhibition of binding or even toxicity
when the same antibody is injected at a later time.
All of these issues can be addressed by the described methods that reduce cross-
20 reactive and nonspecific binding of antibody and antibody conjugates, and are equallyapplicable to any substance that has a nonspecific uptake site or whose receptors are
shared by non-target cells.
The methods of the present invention reduce cross-reactive and/or nonspecific
25 binding of specific antibodies and hormones when a-lmini.~tered to diagnose, stage,
evaluate or treat diseases such as cancer in hllm~n~ The characteristic of specificity
makes antibodies potentially useful agents for targeting defined populations of cells such
as tumor cells that express tumor-specific (expressed uniquely by tumor cells) or tumor-
associated (expressed by tumor cells and by a subpopulation of normal cells) antigens.
30 The clinical utility of these specific antibodies, however, is colllproll.ised by the
phenomenon of cross-reactive and nonspecific binding.
X
- 1340312
s
One method of markedly (limini~hing nonspecific uptake is to remove a nonspecific
binding portion of the antibody, leaving the antigen binding portion (e.g., F(ab)'2,, Fab',
Fab or Fv). Fragments may accumulate more rapidly onto the tumor cell than wholeantibody due to their smaller size, which facilitates egress from the circulation across
blood vessel and capillary walls into the tumor bed. This does not usually compensate for
the decreased scrum half-life of the fragments reslllting in decreased tumor accumulation
compared to while antibody. Thus there exists a need to prolong serum half-life of
fragments in order to take advantage of their more rapid accumulation in target cells and,
therefore, to improve localization into a tumor. Additionally, nonspecific binding of
specific antibodies into normal organs needs to be decreased.
Generally, a major obstacle to the successful clinical use of antibodies and conjugates
thereof has been inadequate delivery to target cells. This has been assessed both by
immunohistochemistry on frozen sections of tumors removed after antibody
~(lmini~tration (Oldham, R.K. Foon, K.A., Morgan, A.C., et al. (1984) J. of Chem. Oncol.
1235-44; Abrams, P.G., Morgan, A.C., Schroff, C.S., (1985) In: Monoclon~l
~ntibollies ~n(l C~ncer Ther~y (Deisfeld and Sell, Eds.), Alan R. Liss Inc., pp. 233-36)
or by calculating the percent of the radiolabeled antibody dose per gram in the tumor
(Murray, J.L., Rosenblum, M.G., Sobol, R.E., et al. (1985) C~ncer Res. 45: 2376-81;
Carrasquillo, J.A., Abrams, P.G., Schroff, R.W., et al. (submitted); Epenetos, A.A.,
Mather, S. Gwanowska, M., et al. (1982) Lancet II: 999-1004; Larson, S.M.,
Carrasquillo, J.A., Krohn, K.A., et al. (1983) J. Clin. Tnvesti~tion 72:2101 -2114 (1983).
There was clear evidence in the former studies of increased antibody localization in
tumor with higher doses of specific antibody. The quantitative studies with radiolabeled
antibody have shown 0.0001%-0.004% of the injected dose/gram loc~li7ing to tumor(Carrasquillo, iki~.).
Radiolabeling of antibodies permits the quantitative assessment of percent accretion
into tumors and normal organs and disappearance from the blood and the whole body.
This serves as a paradigm for the biodistribution and kinetics of antibodies andimmunoconjugates. Conjugated or labeled peptides or steroid hormones provide an
analogous strategy.
X
13~031~
-6-
Studies with radiolabeled antibodies have demonstrated that part of the problem
causing low tumor accumulation is the localization of radiolabeled antibody in other
organs, such as liver, spleen, narrow, lung or kidney. The prior art is replete with
examples of increasing mass of specific antibody to improve tumor localization (e.g.,
5 Abrams, iki~; Murray, iki~; Carrasquillo, iki~; Epenetos, i~i~; Larson, i~ ). The present
invention uses irrelevant antibody to absorb to these nonspecific sites, thus obviating the
requirement for large doses of unconjugated specific antibody required in the prior art.
One advantage of this strategy is that unconjugated irrelevant antibody does not compete
for specific sites on target cells with conjugates specific antibody. Another advantage is
10 that irrelevant antibody skew the immunological response of the patient away from the
specific antibody (vide infra).
One measure of decreased adsorption of specific antibody onto nonspecific sites is
prolonged serum half-life ofthe antibody. Another unexpected result of pre-allmini~tering
15 isotype and subclass matched whole irrelevant immunoglobulin is the prolonged serum
half-life, reduced nonspecific adsorption and improved tumor detection with fragments
of specific antibody.
The prior art recognizes the problem of loc~li7.ing specific antibody and fragments
20 and conjugates thereof due to nonspecific and cross-reactive binding to non-target
epitopes, e.g., Murray, iki~. Yet, the solution employed has been to lump together the
problems of nonspecific and cross-reactive uptake and to use a single strategy -- co-
a-lmini~tration of large masses of specific, unconjegated antibody -- to overcome the
problems. The present invention not only distinguishes these problems, but also offers
25 different solutions to each: (1) the use of irrelevant antibody (immunoglobulin) to reduce
nonspecific and cross-reactive binding of specific antibody, and (2) the a-lmini.~tration
of unconjugated specific antibody to bind to cross-reactive sites prior to the
a-lmini~tration of conjugated specific antibody. The latter solution applies equally to
steroid and peptide hormones.
Since the antibodies are proteins, in most cases of non-human origin, the spectre of
human antiglobulin and anti-idotype has been raised (Oldham, il~i~; Abrams, iki~;
134o~l2
-7-
Murray, i~i~; Carrasquillo, i~i~; Epenstos, i~i~; Larson iki~.). This invention
demonstrates that the co-administration of larger masses of irrelevant antibody can skew
that antiglobulin response toward the irrelevant antibody and not the specific antibody
so that, along with either the same or a second irrelevant antibody, the same target-
5 specific antibody can be injected again and localize in tumor sites without significantformation of antiglobulins to the specific antibody.
Despite their increased specifity, monoclonal antibodies thus far have not been
entirely tumor-specific, but rather recognize tumor-associated antigens. When their cross-
10 reactivity is predomin~ntly to one organ that can be selectively perfused, the presentinvention provides methods for targeting specific antibody conjugates by selective
directive infusion of unconjugated specific antibody into a vessel feeding a normal organ
known to concentrate conjugated antibody by cross-reactive binding. For a therapeutic
conjugate, the reduction in toxicity is a substantial advantage.
Conjugates of specific antibody may localize in nonspecific sites due to the molecule
linked to the antibody rather than the antibody itself. The protein toxins (e.g., ricin, abrin,
diphtheria toxin, pseudomonas toxin) can bind to m~mm~ n cells through particular
portions of these molecules. This property has prompted a large effort to elimin~te their
20 nonspecific binding capability while preserving the potency of the toxin. According to
the methods of the present invention, detoxified protein toxins, when conjugated to
nonspecific antibody, can reduce binding of the specific antibody-toxin conjugate to
nonspecific sites. This methodology reduces whole organism toxicity and permits the
~(lmini.~tration of larger doses of the specific toxin conjugate.
ni~closure ofthe Tnvention
The method of the present invention is for enhancement of delivery to target cells of
antibodies or fragments thereof or other receptor-mediated delivery systems, such as
30 peptide, specific for a population of cells of a m~mm~l. The method comprises the steps
of ~(lmini.ctering to said m~mm~l an adequate dosage of blocking antibodies or fr~gment~
thereof or other receptor-mediated delivery systems, such as peptide, and ~-lmini~tering
X
13403t2
-8-
to said m~mm~l an effective dosage of said antibodies or fragments thereof or other
agents that target a defined population of cells via receptors, such as hormones, specific
for said population of cells, the blocking antibodies or fragments thereof or other
targeting agents capable of nonspecific and/or cross-reactive binding to non-target cells.
5 The antibody fragments of either the blocking antibodies or the specific antibodies are
selected from the group consisting of F(ab)', F(ab')2, Fab, Fv and mixtures thereof.
In another aspect of this invention, there is provided a method for enhancing targeting
of antibodies or fragments thereof to target cells in a cell population which comprises
10 communicating with the cell population blocking antibodies or fragments thereof and the
antibodies or fragments thereof which are specific for said target cells.
In a preferred embodiment, the target cells are characterized by having tumor-
associated antigen. Both the blocking antibodies and the specific antibodies may be either
15 monoclonal or polyclonal antibodies. Furthermore, the specific antibodies and fragments
thereof may be conjugated to cytotoxins, radionuclides or biological response modifiers.
The ~flmini~tration of the blocking antibodies is preferably done prior to the
administration of the specific antibodies; alternatively, such blocking antibodies may be
20 a(lmini~tered simultaneously with the specific antibody. The effective dosage of the
antibodies or fragments thereof specific for a said population of cells is either
diagnostically effective or therapeutically effective. The preferred m~mm~l of the
methods disclosed herein is man.
An additional preferred embodiment of the methods disclosed herein is for
enhancement of the localization of antibodies or fragments thereof specific for a cross-
reactive antigen contained within a m~mm~l's tissue or organ. This method comprises
the steps of directly perfusing said tissue or organ with an adequate dosage of blocking
antibodies or fragments thereof and then ~-lmini~tering systemically to the m~mm~l an
effective dosage of said antibodies or fragments thereof specific for said antigen
contained within said tissue or organ that is also present on target cells.
X
1340312
g
A related aspect of the present invention discloses a method for reducing within a
m~mm~l the production of anti-immllnoglobulin directed against antibodies or fragments
thereof specific for a population of target cells. This method comprises the steps of
a~mini.~tering to said m~mm~l an adequate dosage of blocking antibodies or fragments
5 thereof capable of stimulating the production of anti-immunoglobulin directed against
said blocking antibodies or fragments thereof and ~-lmini~tering to said m~mm:~l a
therapeutically effective dosage of said antibodies or fragments thereof specific for said
target cells, wherein said therapeutically effective dosage is smaller than said adequate
dosage of blocking antibodies or fragments thereof.
1. Reducing Non~recific Upt~ke
An irrelevant antibody may be matched with a specific antibody by species, isotype
and/or subclass and is ~(lmini~tered, usually in several-fold or greater quantities, to
15 patients prior to the injection of the specific antibody or specific antibody conjugates.
Alternatively, such matching may not be necessary. The prior art has demonstrated
improved tumor localization of whole antibody with increasing doses of specific
antibody. These results were assessed with either trace-labeled specific antibody or by
immunohistochemistry following antibody ~rlmini.~tration (Oldham, i~i~; Abrams, iki~;
20 Murray, iki~; Carrasquillo, iki~.). In the former, the unlabeled, specific antibody was
usually given simultaneously with the labeled preparation. The underlying concept
involved providing sufficient mass of antibody to attach to nonspecific sites, leaving
more specific antibody available to bind to tumor. There was some reduction of labeled
antibody in nonspecific sites (e.g., liver), and a marked increase in serum half-life of the
25 radiolabeled whole immunoglobulin (Carrasquillo, i~i~.). There is, however, the
potential for competition for specific binding at the tumor site by the labeled and
unlabeled antibody. If the label is used symbolically to represent any agent conjugated
to antibody, competition of unlabeled antibody for the tumor antigen at the mostaccessible (perivascular) sites will colnl)lolllise the delivery of that agent to the tumor.
30 In addition, a percentage of the labeled antibody, when injected simultaneously, behaves
as a true tracer and deposits in normal nonspecific and cross-reactive sties.
1340312
-10-
The method described herein uses irrelevant antibody (blocking antibody) to reduce
the nonspecific uptake, without competing for the tumor antigen, preferably ~-lmini~tered
prior to specific antibody (see Example I). The positive results observed by pre-
~(lmini~tering irrelevant antibody were not expected since human g~mm~globulins, in
5 large (>1 gm) quantities, failed to show any decreased localization of radiolabeled
antibody into nonspecific sites (Carrasquillo, il~i~.).
Further, the method described herein is applicable to fragments of specific antibody
as well as to whole specific antibody. The method also has been shown to increase the
10 serum half-life of fragments.
Particularly unexpected was the prolonged serum half-life in the initial phase of
serum clearance achieved by using whole irrelevant antibody prior to specific antibody
fragments. This was surprising since it was assumed that the nonspecific binding sites on
15 antibody would be elimin~te~l by creating fragments devoid of the Fc portion of the
molecule. This method is useful for reducing nonspecific binding of specific antibody,
whole or fragment, or conjugates of such antibody, whole or fragment, with
radionuclides, drugs, toxins biologic response modifiers or differenti~ting agents.
2. Reducing Cross-Reactivity with Norm~l Ti~lles.
Normal tissues may contain an epitope, or structural homolog of an epitope,
expressed on the target cell. Several methods to reduce localization to non-target tissues
are presented herein. For normal tissues equally or more accessible than tumors to
25 specific antibody, unconjugated specific antibody (blocking antibody) is ~lmini~tered
prior to conjugated specific antibody (see Example II). The antigen sites on normal
tissues, such as blood vessels or liver, will first bind the previously injected,
unconjugated whole antibody. In one embodiment, the blocking antibody is bivalent
(whole or F(ab)'2) and the conjugated antibody is monovalent (Fab', Fab or Fv). This
30 approach takes advantage ofthe higher antigen affinity ofthe bivalent molecules and thus
reduces the competition ofthe subsequently a-lmini~tered, conjugated monovalent species
for the cross-reactive sites.
~ 134031~
-11-
The precise doses of the cold-specific antibody and the timing of a(lmini~tration will
be dependent upon the size of the patient, the quantitative normal tissue expression of the
antigen or epitope, the relative accessibility of tumor compared to normal tissue sites, and
the purpose for which the entire procedure is being performed. For example, a modest
5 degree of nontumor targeting may be acceptable in a therapeutic study so long as toxicity
is tolerable, but less acceptable in an im~ging mode where false positives could be a
problem.
The dose of the cold-specific antibody can be approximated by a-lmini~tering
10 increasing doses of the cold-specific antibody prior to biodistribution studies with a
labeled antibody or labeled conjugate. When normal tissue sites that show evidence of
uptake of the radiolabeled antibody or conjugate with no cold-specific antibody are no
longer vi~ 1i7ed, then an adequate dose of the cold-specific antibody will have been
reached. Alternatively, a greater interval may be allowed between the given dose of cold-
15 specific antibody and the subsequent a-lmini.ctration of conjugated specific antibody to
permit greater permeation of the cold-specific antibody into normal tissues.
Another approach is to determine the serum half-life of an irrelevant antibody or
fragment and pre-a-1mini~ter sufficient cold-specific antibody or fragment to raise the
20 half-life of the conjugated antibody or fragment close to the half-life of the irrelevant
antibody. This would be indirect evidence that a normal tissue antigen "sink" was
completely or largely saturated.
An additional method of det~rmining the a~plopl;ate dose has been described by Eger
25 et al. Using data obtained from patients receiving the 9.2.27 antibody, they constructed
a m~tl~em~tical model that, by regression analysis, could predict a dose of antibody that
would saturate the "antigen sink" in normal tissues.
Once determined for a particular antigen-antibody system, the dose and schedule can
30 be calculated on a body surface area or weight basis for that particular antibody.
1340312
-12-
Within a wide range, however, the ~mini~tration of more than a normal tissue
saturating dose will have little effect on the delivery of the conjugated antibody to the
tumor. The major considerations are saturation of the more accessible tumor sites and
affinity differences between unconjugated, "cold" antibody and the conjugate. If the
5 conjugate has substantially less affinity, too large a dose of unconjugated antibody will
compete off the conjugate and thus efficacy - therapeutic or detection - will be reduced.
A second method for reducing cross-reactivity with normal tissues that contain the
cross-reactive associated antigen involves directed prior or co-~-lmini~tration of
10 unconjugated antibody (see Example III). Generally, venous or arterial access is achieved
through percutaneous catheterization. For example, the renal arteries or veins may be
catheterized to perfuse the kidneys selectively. Alternatively, a catheter may be placed
into an a~plopliate vessel (e.g., hepatic artery for liver) at the time of surgery. The
unconjugated antibody is injected by the catheter to contact the antigen sites of the
15 normal organ with the "first pass" of antibody. The antibody may be whole or
fragmented. Either simultaneously or thereafter, the conjugated antibody is injected into
the m~mm~l's general circulation. Binding of the specific conjugated antibody to the
previously perfused sites in the normal organs is reduced, thus increasing the availability
of the conjugated specific antibody for the target cells. This also results in reduced
20 toxicity to the cross-reactive, non-target organ by the specific conjugated antibody.
A third method involves either peripheral vein or directed infusion as describedabove. Instead of using unconjugated antibody as the blocking antibody, a conjugate of
antibody and detoxified cytotoxin or BRM is infused where the detoxification (i.e.
25 reduced toxicity) process preserves the sites of recognition and/or binding by normal cells
(see Example IV). In a plerelled embodiment, the detoxified agent is either free or bound
to the nonspecific antibody, and the detoxified conjugate is injected prior to the
~mini~tration of specific antibody conjugate and, if necessary, may be ~1mini~tered as
a constant infusion.
In another method, where the tumor to be treated is localized in an organ or a limb,
conjugated antibody is injected via directed catheter into the organ or limb following
X
13~031~
-13-
peripheral blockage of cross-reactive and nonspecific sites with unconjugated specific
and/or detoxified, conjugated irrelevant and/or unconjugated irrelevant antibody.
An additional method provides for injection of unconjugated specific antibody via
5 catheter or other means of directing the first ~(lmini~tration of the antibody to a defined
organ that contains the cross-reactive antigen where the tumor to be treated is
dissemin~te~l outside of the organ. Unconjugated irrelevant antibody and/or detoxified,
conjugated, nonspecific antibody are injected by peripheral vein. Conjugated specific
antibody is then injected intravenously.
3. Reducirlg ~ntiglobulin Respon~es nirected a the Specific ~ntibody.
Irrelevant antibody is ~-lmini~tered prior to or simultaneously with speclfic antibody
or specific antibody conjugate. The irrelevant antibody may be ~-lmini~tered in higher
15 doses than the specific antibody. In one preferred embodiment, the irrelevant antibody
is whole immunoglobulin and the specific antibody is an antibody fragment (see Example
V). In another pl~r~lled embodiment, the ratio of specific to nonspecific antibody ranges
from about 1:1 to about 1:100 and is preferably 1:5. In another preferred embodiment,
the nonspecific antibody is whole immunoglobulin and the specific antibody is Fab or Fv.
20 The basis for this strategy is that the higher dose and use of whole, not fragment, of the
irrelevant antibody is more likely to evoke an immunological response than the lower
dose and less immunogenic fragments of specific antibody.
When a patient develops antiglobulins that bind to the irrelevant antibody but not the
25 specific antibody, the same irrelevant antibody may be ~mini~tered in a subsequent dose
ahead of specific antibody to adsorb the circulating antiglobulin and in higher doses to
block nonspecific sites in normal tissues. Any antiglobulins to irrelevant antibody that
cross-react with specific antibody would be completed and deposited in the kidney or
reticuloendothelial system, depending on the size of the complex. Most importantly,
30 conjugated specific antibody given subsequently would not be completed, but would be
free to bind to target cells. Alternatively, a second irrelevant antibody that is not
X
1~03 L~
-14-
recognized by the antiglobulin response may be substituted in subsequent injections. A
combination of irrelevant antibodies may also be used.
FXAMPI F.S
I. Reducing Non~pecific Upt~ke
Monoclonal antibody (MAb) NR-2AD is a murine IgG2a immunoglobulin that was
designed as an anti-idiotype that bound to a single patient's B-cell lymphoma and to no
other human tissue MAb 9.2.27 is a murine IgG2a antibody that recognizes the 250Kilogalton glycoprotein/proteoglycan melanoma-associated antigen. Both were scaled
up by in vitro cell culture, purified by column chromatography, and tested for purity and
sterility to meet the draft guidelines for injectable monoclonal antibodies from the Of fice
of Biologics, Food & Drug Administration ("Points to consider in the m~mlf~cture of
injectable monoclonal antibody products intended for human use in vivo: Revised draft
of 6/11/84"). MAb 9.2.27 was digested with pepsin and the F(ab)'2 fragment purified
from residual intact antibody.
NR-2AD 50 mg was diluted in normal saline and injected intravenously into patient
8501.08 who had metastatic malignant melanoma. One hour later,2.5 mg Tc-99m labeled
9.2.27 F(ab)'2 was ~-lmini~tered intravenously. The patient's blood was withdrawn and
counted with a Tc-99m standard. Surprisingly, the serum half-life (tl/2) of the labeled
F(ab)'2 was 17 hours. The average serum tl/2 of the labeled 9.2.27 F(ab~' for patients
receiving NR-2AD was 12 hours compared to 6 hours for those who did not receive NR-
2AD prior to the labeled fragment. The patient's tumor was vi~ li7ed by gamma camera
im~ging and then excised. Another unexpected result was the detection of a tumor that
weighed only 0.25 gm.
II. Reducing Cros~-Reactive Rin~ing of Specific Antibody Conjl~t~
A. Cold-specific ~ntibody block~ upt~ke of conjl~t~-l specific ~ntibody in norm~l
nrg~
The effect of reducing cross-reactive uptake of a radiolabeled monoclonal antibody
p~ ion was ~sessed in the context of diagnostic im~ging clinical trial. The patient
(#8501.22) examined in the study was a 32-year-old male with a metastatic melanoma
X
13~0312
-15-
lesion in the left posterior portion of the neck. The lesion consisted of a tumor involved
lymph node measuring 2x2 cm.
The patient received two schedules of radiolabeled antibody three days apart. The
5 first schedule of antibody consisted of two doses: a 50 mg dose of non-radiolabeled intact
irrelevant antibody (NR-2AD) ~flmini~tered intravenously followed 1 hour later by a 2.5
mg dose of Tc-99m radiolabeled Fab fragment of MAb 9.2.27. The localization of
radiolabel within the patient was assessed by gamma camera im~ging over a 7 hourperiod. The second schedule was identical, except that 7.5 mg of non-radiolabeled F(ab)'2
fragment of 9.2.27 was ~mini~tered 5 minlltes prior to the radiolabeled Fab preparation.
The 50 mg dose of intact irrelevant antibody was aAmini~tered in both schedules, in order
to reduce nonspecific uptake of the radiolabeled antibody into normal tissues. The non-
radiolabeled target specific F(ab)'2 preparation was ~-lmini~tered for the purpose of
reducing cross-reactive uptake of the radiolabeled antibody by non-target tissues.
The results of the study showed that omission of the non-radiolabeled target-specific
antibody (9.2.27 F(ab)'2 was accompanied at 7 hours post-infusion by uptake of
radiolabeled antibody into non-target tissue, specifically spleen, bone marrow and kidney.
The known tumor site was not imaged. Pre-infusion of the non-radiolabelèd target-
specific antibody (9.2.27 F(ab)'2 was accompanied at the same time point by uptake in
the kidney, but no demonstrable uptake in the spleen, bone marrow or other normal
organs. This result was unexpected since the marrow and spleen uptake had previously
been assumed to be nonspecific. In addition, the know tumor in the neck was clearly
visible, confirming that blocking cross-reactive binding sites by prior infusion of non-
radiolabeled target-specific antibody both reduced non-target tissue localization and
increased tumor localization of the radiolabeled target-specific antibody.
B. Cold-Specific Antihody Rlock~ Upt~ke ofthe T ~heled Specific Antibody by
~n Fpitope-Specific Mech~ni~m.
Patient 8501.29 was a 49-year-old C~llc~ n male who presented with melanoma on
his back that then spread to lymph nodes under his arm. At the time of im~ging with the
1340312
..i
-16-
99mTc-labeled antibodies, the patient had presumed disease in the right axilia (recurrent),
questionable small subcutaneous metastases on his arm and back, and diffuse hepatic
involvement, as evidenced by CT scan and elevated hepatic tr~n~min~es.
The purpose of the test in this patient was to determine if the cold-specific antibody
blocked normal tissue accumulation of the labeled antibody in an antigenbinding site
(i.e., epitope-specific) manner. Two antibodies, MR-ML-05 and 9.2.27, that each
recognized distinct epitopes of the 250 Kd. glycoprotein/proteoglycan melanoma-
associated antigen were chosen for this study.
For the first procedure, the patient received irrelevant antibody NR-2AD (Nl~
900.00), cold-specific 9.2.27 (NRX-112), and then radiolabeled NR-ML-05 (NRX
118.03) so that the cold-specific blocker and the labeled antibody recognized distinct
epitopes of the same antigen. Gamma camera im~ging 4 and 8 hours after injectionrevealed hepatic metastases, an auxiliary chain of nodes, but also images of marrow
(sternum and pelvis) and spleen. A repeat procedure was identical but substituted MR-
ML-05 (NRX 118) as the cold-specific blocker for the 9.2.27 used in the first procedure.
Gamma camera im~ging revealed the same sites of disease, but this time with the absence
of marrow (sternum and pelvis) and splenic uptake of the label.
This example, using the same patient as his own control, demonstrated conclusively
that the cold-specific antibody blocked uptake of the labeled specific antibody into
normal organs in an epitope-specific manner.
III. T)irected Tnfil~ion to Reduce Cross-Reactive ~inlling
Percutaneous catheterization of the celiac and/or pancreatoduodenal arteries is
performed through the femoral artery. MAb NR-CO-l that reacts with colon carcinoma
and cross-reacts with normal pancreas is labeled with Tc-99m. Unlabeled NR-CO-l
F(ab)'2 fragments (10 mg) are injected by 10 minute infusion through the catheter
followed by a normal saline flush (directed infusion). At the end of the directed infusion,
1340312
-17-
the labeled NR-CO-l (2.5 mg) is injected by peripheral vein. Omitting the directed
infusion results in increased localization in the pancreas of the subsequently
a~mini~tered, labeled antibody, indicating that the directed infusion can selectively
reduce binding of conjugates in organs containing cross-reactive antigen.
s
IV. Reducing Non.~recific Rin(ling of Conj~ t~.
Pseudomonas exotoxin (PE) and diphtheria toxin (DT) are bacterial proteins that
10 interfere with protein synthesis in cells by ADP-ribosylation of elongation factor-2 (EF-
2). DT binds to nicotinamide at glutamic acid in position 148 in the molecule.
Substitution of aspartic acid in this position (DTASp) reduces the activity of DT to 1 % of
the native molecule, but does not affect its binding to cells. PE is thought to have a
"pocket" for binding to its substrate that contains a tyrosine (position 481) and glutamic
acid (position 553). Iodination of tyrosine 481 (PE~ ) is expected to result in >50% loss
in activity of PE (PE~l) but retention of its binding ability.
NR-2AD (irrelevant immunoglobulin) is reacted with 25 mM dithiothreitol and PE~Ior DTASp with SMPB [succinimidyl (4-p-maleimidopheny 1) butyrate], and the unreacted
20 reagents are removed. The derivatized antibody and toxin are mixed together, producing
NR-2AD-S-C-PE~l or NR-2AD-S-C-DTASp (detoxified conjugates) that are purified bycolumn chromatography. 20-500 mg of the detoxified conjugates is ~tlmini~tered
intravenously prior to a-1mini~tration of NR-ML-05-S-C-PE or NR-ML-05-S-C-DT
(specific toxin conjugates). NR-ML-05 is a murine monoclonal antibody that recognizes
25 a 250 kilogalton glycoprotein/proteoglycan human melanoma-associated antigen.
The prior a-lmini.~tration of the detoxified conjugates is expected to reduce specific
conjugate binding in normal nonspecific sites (nonspecific has two components in this
example, viz, the nonspecific binding of the irrelevant immunoglobulin part of the
30 conjugate and the nonspecific binding of the modified toxin). The specific toxin
conjugate is then injected intravenously in doses two times or greater than could be given
without the blockade of non-specific sites with detoxified conjugates. Additionally,
1,~4031~
-18-
unconjugated NR-2AD and unconjugated NR-ML-05 are given prior to the specific toxin
conjugates. These methods permit the a-lministration of higher doses of potent
immunoconjugate for increased localization and improved tumor reduction.
V. Reducir~g Antigloblllin Respon~ nirected at a Specific Antibody
The effect of co-~-lmini~tration of an irrelevant monoclonal antibody (NR-2AD) on
subsequent antiglobulin development to a specific MAb (9.2.27) was assessed in the
context of a diagnostic im~ging trial. Sixteen patients with metastatic malignant
melanoma received 1 to 25 mg doses of Tc-99m radiolabeled 9.2.27 monoclonal antibody
for fragments thereof, preceded by 5 minlltes by doses of 7.5 to 9 mg of non-radiolabeled
9.2.27 or fragments thereof. Thus, patients received total doses of 10 mg of the 9.2.27
antibody and its fragments. One hour prior to a-lministration of the specific antibody, 50
mg doses of a non-radiolabeled intact irrelevant antibody (NR-2AD) was a-lministered
to 11 of the 16 patients.
Serologic evidence of an antiglobulin response was evaluated using a solid-phase,
enzyme-linked immunoassay (ELISA) employing either the target-specific (9.2.27) or
irrelevant (NR-2AD) antibody as the capture antigen. Serum specimens were evaluated
from time points ranging from 2 weeks to over 6 months following treatment. Six of the
eleven patients evaluated receiving NR-2AD demonstrated evidence of antiglobulinresponse to the irrelevant NR-2AD antibody of greater magnitude than that which was
obtained in the upper 95th percentile of 60 healthy controls. In the same group of eleven
patients, only one person demonstrated an antiglobulin response to the target-specific
9.2.27 antibody above the 95th percentile of the healthy control population.
These data indicate that prior a lministration of a 5-fold excess of an irrelevant MAb
preparation resulted in antiglobulin responses, in those individuals who developed an
antiglobulin response, which was skewed toward the irrelevant rather than the target-
30 specific MAb.
VI. Repeated Aclministr~tion of Antibody in the Presence of Antiglobnlin.c
X
...
1~031 2
-19-
Patient #8501.08 in the 99mTc im:~ging trial of melanoma received two im~ging
procedures within a one-week period with the 9.2.27 melanoma-specific antibody and the
NR-2AD irrelevant antibody. The patient returned after 8 months and was repeat imaged
5 with the NR-ML-05 melanoma-specific antibody and the NR-2AD antibody. At that
time, the patient had developed a 1 9-fold increase in antiglobulin directed to the NR-2AD
antibody. After premedication with 50 mg diphenhydramine, the patient received NR-
2AD 41 mg diluted in normal saline slowly over 60 minutes. He then received 10 mg
NR-ML-05 cold-specific antibody followed by NR-ML-05 Fab labeled with 99mTc. The10 labeled antibody showed no evidence of altered biodistribution and succec~fi-lly targeted
known subcutaneous and liver tumors.
By contrast, studies in normal guinea pigs immunized with murine monoclonal
antibodies and imaged with either the same or irrelevant 99mTc-labeled antibody
15 demonstrated that the presence of specific antiglobulin altered biodistribution of labeled
antibody. Administration of a radiolabeled antibody to which antiglobulin was reactive
resulted in biodistribution of the radiolabel to liver and spleen. If the circulating
antiglobulin was not specific for the radiolabeled antibody, then biodistribution was
indistinguishable from biodistribution in the non-immunized animal.
From the foregoing, it will be appreciated that, although specific embodiments of the
invention have been described herein for purposes of illustration, various modifications
may be made without deviating from the spirit and scope of the invention. Accordingly,
the invention is not limited except as by the appended claims.
